openerp-community team mailing list archive

["Dale Scott" <dale@xxxxxxxxxxxxx>]

Can anyone comment on real-time integration of OpenERP with existing legacy
systems (instead of using an "import and cut-over" strategy)?

My organization has a number of legacy systems: client job
planning/scheduling, part number management, manufacturing process
management and serialized part tracking, and issue-tracking. Would it be
advisable (and practical) to connect a new OpenERP to this data in real-time
(perhaps think "continuous import") as an iterative implementation strategy?

The users performing data entry could continue to use their familiar legacy
systems while senior management (primarily interested in aggregated data
reports) start using OpenERP reporting and dashboards with current data.
Then as OpenERP features and workflows become better understood (and support
increases from senior management), legacy systems would be replaced with
OpenERP - first the legacy systems under my personal control (I'm the
manufacturing leader), with others coming later as OpenERP knowledge and
acceptance increase. For example, our Operatins team uses the client
planning system to plan and mange client jobs, while I only need the job
number and date for checking out serialized parts (this is currently all
done by re-typing data from one system into another, there is no integration
between the various legacy systems).

I've implemented scheduled (cron) export from the legacy systems to CSV with
import into a MySQL database (using one or more tables in the same database
for each legacy application). I've also created a link table for
establishing relationships between legacy applications (for example, to
relate part numbers in the legacy part number system to corresponding part
numbers in the legacy manufacturing system; part numbers in the two systems
are similar but not identical due to their history). My thought was to use
the MySQL database as an integration point for control, and to view and
validate the legacy data before being imported by OpenERP, but OpenERP could
just as easily import the CSV data directly if that was more efficient or
effective.

Is this reasonable? Possible? I'd really appreciate some specific direction
if anyone has some thoughts.

Dale